Voice recognition systems are known and their use is increasing in popularity. Such systems can be incorporated into interactive systems for carrying out selected automated commands. These systems are increasingly being used in environments that are not solely populated by the user, such as in automobiles. Where multiple user environments exist, a concern arises as to the voice recognition and interactive system performing an unwanted command, either through erroneous recognition of the command or performance of a command issued by an unauthorized user. Unwanted performance of commands can be particularly dangerous in environments such as automobiles.
In a Free-to-Talk voice interactive system, the device is always in an active standby state, and there is no requirement to perform a user-initiated action before issuing a voice command. Input voice signals are always considered by the voice interactive system as voice commands and the likelihood of unwanted commands being performed is increased with noisier environments.
Contemporary voice recognition and interactive systems have sought to address these concerns through activation requirements for operation of the system. Two such voice recognition and interactive systems are Push-to-Talk systems and Talk-to-Talk systems. Push-to-Talk voice interactive systems require a user-initiated physical actuation to activate the system each time before a voice command can be evaluated by the device. Talk-to-Talk voice interactive systems require a keyword to be spoken to activate the system each time before a voice command can be evaluated by the device.
However, neither the Push-to-Talk or Talk-to-Talk voice interactive systems provide any safeguard once the system has been activated. Noise or unauthorized commands subsequent to the actuation steps described above can still result in the performance of unwanted commands by the interactive system. Additionally, such systems can be disadvantageous in certain situations because of the physical actuation being required and can be disadvantageous in other situations because of the physical or verbal initiation step being required prior to each activation of the system.
The Push-to-Talk or Talk-to-Talk voice interactive systems require the physical or verbal actuation steps described-above regardless of the particular command that will be issued by the user. Such an inflexible system is not user friendly and creates inefficiencies in the system.
In other contemporary voice interactive systems, the physical actuation step has been completely eliminated and only a single keyword is required to be spoken to enable the device prior to issuing a series of commands to the device. Such contemporary devices use an idle time technique to determine whether another keyword must be spoken to enable the device prior to issuing more commands to the device. However, these systems still suffer from the drawback of being susceptible to performance of unwanted commands once the single keyword has been spoken. Additionally, the use of the single keyword can be problematic where unauthorized users learn of the keyword. For example, in an automobile where passengers are present, a keyword may be repeated by passengers, e.g., a child, resulting in the potential of unwanted performance of automated commands by the interactive system.
A need therefore exists for a system and method for voice interactive systems wherein the above-mentioned disadvantages may be alleviated. A further need exists for such a system that provides sufficient safeguards against unwanted performance of commands while maintaining the efficiency of the system. An additional need exists for such a system that can apply varying levels of safeguards based upon the particular command being issued. An efficient, safe and secure voice interactive system is necessary to encourage adoption of such systems in pervasive environments.